Coordinated PSO-PID based longitudinal control with LPV-MPC based lateral control for autonomous vehicles

Kebbati, Yassine; Ait-Oufroukh, Naima; Vigneron, Vincent; Ichalal, Dalil · 2022 · Crossref

DOI: 10.23919/ecc55457.2022.9838192

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of coordinated longitudinal and lateral control for autonomous vehicles, aiming to improve tracking accuracy and ride comfort under varying driving conditions. The authors highlight that longitudinal dynamics significantly influence lateral stability, yet many existing solutions treat these controls separately or rely on computationally expensive nonlinear models. To bridge this gap, the study proposes a hybrid control strategy combining an improved Particle Swarm Optimization-tuned PID (PSO-PID) controller for longitudinal speed tracking and an enhanced Linear Parameter Varying Model Predictive Control (LPV-MPC) for lateral trajectory tracking. The methodology involves modeling the vehicle’s longitudinal dynamics using a nonlinear approach that accounts for powertrain and tire characteristics, while lateral dynamics are represented by an LPV bicycle model. A key innovation is the use of a Recursive Least Squares (RLS) estimator to online-estimate tire cornering stiffness coefficients, allowing the LPV-MPC prediction model to adapt iteratively to changing road conditions. For longitudinal control, the authors develop an improved PSO algorithm with dynamic inertia weight and acceleration coefficients to optimize PID gains, using Mean Squared Error as the fitness function. The lateral controller employs an enhanced cost function featuring a relaxation factor to prevent constraint infeasibility and an exponential weight to prioritize immediate error correction. The system was validated through Matlab/Carsim co-simulations, with the LPV-MPC solver running at 20Hz. The results demonstrate that the proposed coordinated strategy achieves superior performance compared to standard approaches. The PSO-PID controller effectively handled speed tracking tasks, including scenarios with wind disturbances, by dynamically switching between acceleration and braking commands. The enhanced LPV-MPC showed significant improvements in lateral tracking accuracy and smoothness. Specifically, the enhanced MPC achieved a Mean Squared Error (MSE) of 2.118e−4 for position tracking and 1.228e−4 for orientation tracking, vastly outperforming the standard MPC, which recorded MSEs of 0.601 and 6.756e−4, respectively. The adaptive nature of the controller allowed it to maintain stability and precision even when tire parameters varied, ensuring smoother steering inputs and better ride comfort. The significance of this work lies in its ability to provide a computationally efficient yet highly accurate control framework for autonomous driving. By integrating online parameter estimation with adaptive optimization techniques, the approach overcomes the limitations of fixed-gain controllers and the high computational costs of nonlinear MPC. The findings suggest that this coordinated control strategy is robust against parametric uncertainties and external disturbances, offering a viable solution for real-time implementation in advanced driver-assistance systems and autonomous vehicles.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• lane positioning